Pavilion 15 won't boot

[LucaLauro]

Hi,
this is not a support request, it's the follow up of this topic: Pavilion 15 won't boot - 9192553 .
I'm posting this update hoping it might be helpful to people who have had the same problem.

After a looong journey, I managed to understand and fix the problem (spoiler: it was a hardware problem); in fact I'm literally writing this post with my fixed pc right now.

after trying every possible solution that did not involve hardware:
• reset motherboard controller by holding down power button for 30 seconds;
• force BIOS recovery from mass storage by turning on the computer with pressed key combinations such as WIN+B / WIN+V (which by the way in recent models will happen automatically after failed BIOS checksum);
• perform BIOS recovery from USB recovery flash drive created using HP BIOS Update Utility...

since the warranty had been expired for long and an original motherboard replacement from HP would have cost me ~700 € + service, I took it to 3 repair centers in my area.
They all told me that after diagnostics, no obvious problems emerged, and a thorough investigation would cost more than a new PC.
They were right, in fact, motherboard did not present any visible damage nor blown component.
At that point I decided to try fix it myself; so I looked for the schematics of the most similar motherboard model possible (in my case it was a G7HA produced by Quanta) and started to do an in-depth diagnostics of motherboard components.

Basically, that's what happened: 
the 12V surge from external microcontroller shorted the power switch of the USB port it was connected to and enden up in the 5V power rail (🥲).
From there it damaged the APU-CPU Voltage Regulation Module (+VDD_CORE, +VDDA_CORE power rails) gate-driver and, consequently, APU-CPU VRM MOSFETS causing the surge to pass through APU-CPU (🥲); and audio CODEC too.
luckily, being a component designed to work with high currents, APU-CPU was not damaged (the gate driver probably failed to protect it because the surge entered from the control logic power line).

I bought necessary tools and spare parts from Aliexpress for a total of ~100€ (mostly the tools), replaced them and guess what? after the surgery it turned on.
But then another problem arose:
the pc would turn on very rarely (2-3 times in a week); I couldn't turn it off or I risked not being able to turn it on when I needed it. Most of the times I tried to turn it on from completely off, the display would remain completely black; but sometimes (often when left off for at least a night) it would turn on showing a corrupted bios error and an attempt to recover the bios. So, again, tried all possible software-firmware solutions; nothing.
After that I bought a CH341B programmer, tried updating various BIOS versions (which have been difficult to extract because the latest Windows updates did not run properly or did not allow the HP BIOS Update Utility to run) and still nothing; the behaviour was the same.
But actually, thinking about it, if the pc would sometimes boot, it's almost impossible that it was a bios corruption problem or a damaged BIOS flash memory.

I took a good look at the flash datasheet on my motherboard and noticed that in the configuration register (where you set the bits for write protection and other functions) some bits allowed you to set the SPI communication mode with the APU: normal/dual/quad/QPI.
I checked the behavior of the pins when I tried to boot the board and it was that of QPI mode, but the flash was configured for normal SPI. So probably, since the pc was on when the overvoltage entered the APU-CPU, this "corrupted" the configuration register of the flash BIOS by reconfiguring the communication mode (it is enough to change the state of a bit).
So, in a nutshell, the APU was talking in QPI mode while the flash was in normal SPI mode; this would also explain the failure of the bios recovery attempts.
(it should also be noted that the CH341A/B programmer automatically seems to put the memory in normal SPI mode every time it connects and performs the flash chip recognition).

In conclusion, I reloaded the bios into the flash and set it to the correct mode.
From there the pc works perfectly.

I re-entered all the original data in the UEFI (serial number, various information...) since it had been rewritten, even if not strictly necessary, and now it works fine; only two errors are displayed when starting from completely off:
502 - "the CMOS checksum is invalid..."
and
90D - "The computer shut down previously to avoid overheating..."
which have no reason to appear, I think they are two flags created in the EFI partition when the accident occurred and should disappear with a normal proper BIOS update (which I can't do because updated Windows does not run HP Bios Update Utility🙄).

I apologize for the length of this post, hope I haven't bothered anyone and above all I hope it can be useful and save time for those who run into similar problems.

Thanks for your attention and wish you a good time.